As described in Chapter 2, the pastoral areas of tropical Africa are characterised by rainy seasons of short duration, which may be unimodal or bimodal in form, during which insufficient rainfall is received to sustain arable agriculture. Although 'mixed farming'- in the temperate latitude sense - is practiced in some of the more climatically favoured subtropical countries, the climatic control has tended to influence indigenous people to farm with either crops or livestock and rarely both (Philips, 1966). Thus, an ecological adjustment to the climatic environment has become 'fixed' and traditional in the course of time.
Increasing pressure from population growth has caused the replacement of relatively high-quality grasslands by crops, especially since the development of drought-resistant crop varieties; so much so that it can now be accepted that the pastoral areas are confined to those climatic zones which suffer from a seasonal deficit of rainfall of varying severity.
A prerequisite for full development of the pastoral areas, therefore, is the provision of additional and well distributed water supplies. New watering points must be sited in relation to the location and carrying capacities of available pastures (Webster and Wilson, 1966). They must be accompanied by stock control and a system of rotational grazing to prevent overgrazing. This is a management problem, of course, but the water engineer must not be allowed to make an existing situation worse, by the unplanned provision of watering points which will themselves encourage overstocking and overgrazing.
Further, it is clear from Chapter 3 that the interdependence of the components of the hydrological cycle weighs against any uncoordinated action, such as the sole provision of water supplies, since those very supplies may be dependent ultimately upon the maintenance of good infiltration rates. These in turn are a function of vegetation cover, and that cover will only remain if overgrazing is prevented.
Surface water supplies are more prone to degrade and decline due to overgrazing than are groundwater supplies because, generally speaking, stock watering requires so little water compared to the volume of recharge of groundwater. Surface water supplies, however, are usually cheaper to maintain and can be constructed without the aid of sophisticated machinery. Choice of the type of supply will depend partly on the resource potential and partly on the availability of manpower, machinery and funds for construction and maintenance. Before going on to discuss methods of using the water resource, it is helpful to consider the types of resource and their characteristics.
Natural springs occur wherever the hydraulic head in an aquifer exceeds the surface elevation. Often this results in seepage around discharge areas which is revealed by vegetation changes, particularly the occurrence of phreatophytes. Where more open conduits are found springs result, often of remarkably constant discharge and of high quality. On the other hand, they can be highly mineralized, and the vast majority of springs, particularly in semi-arid areas, have discharges less than 1 l/sec. Areas of volcanic rock and sand hills are noted for their potential to contain springs of nearly constant discharge. A classic example of a large spring of high discharge is the complex at Mzima Springs in Kenya, where recharge on the volcanic Chyulu Hills is translated into a flow large enough to serve the town of Mombassa.
Basal springs from inselbergs or monad-nocks, i.e. erosion residuals on the peneplain surfaces of the African continent, have been a source of water for pastoralists for many generations. Only recently has there been a serious threat to their survival as encroachment into the forested upper slopes, felling and burning have laid open the hitherto protected soils to erosion. Infiltration and recharge are the first parameters to change, but there are also more subtle effects, such as the cessation of occult precipitation and 'fog drip'.
The effects of forest removal on the survival of basal springs is a complex and controversial subject. Many people believe that forests "attract' rainfall, with the corollary that removal of the tree cover leads to a decrease in rainfall. Attractive though this thesis is for foresters and conservationists, there appears to be little indisputable scientific evidence to support it (Penman, 1963). The dominant rainfall mechanisms are features of the general circulation but, as was discussed in Chapter 2, in the continental tropics, away from the direct influence of the trade winds, uniformity of pressure and temperature characterises large areas. Within these areas small differences in the surface energy balance may be sufficient to trigger the rainfall producing mechanism, particularly when unstable, moisture-laden air is passing over cool, transpiring areas of forest.
When the forests occur in upland areas, it is difficult to separate the effects of orographic uplift from any cooling effect due to the latent heat of vaporisation required by the transpiration process. Observable effects of forest are the increase in occult precipitation and fog drip (see Section 2.8). It is also irrefutable that forests have a vital role to play in soil and water conservation. Maintaining a forest cover ensures that recharge to groundwater takes place, although this has to be balanced against evaporation and transpiration from deep-rooted trees.
The distinction between a stream and a river is arbitrary both in the scientific and ordinary usage of the terms. A river is a large flow of water and a stream is a small flow, the smallest streams being the beginning of channel flow. River is also used in a generic sense, meaning a system of streams within a river or drainage basin. Both may be seasonal or perennial. Even the perennial streams and rivers originating in pastoral areas are characterised by extreme variability in frequency and volume of discharge.
In this context a stream may be defined as a natural channel flow where the catchment or drainage basin is less than 10 km2. Peak runoffs could be expected to lie in the range of 10 to 200 m3/sec-1, depending on the type of soil, the slope, the degree of vegetation protection and the shape of the catchment. The larger floods would be exceptional and confined to steep, rocky and unprotected catchments. They are included here as a reminder of the immense destructive power of floods from even quite small catchments which are degraded and eroded.
Annual yield can be expected to be less than 100 mm depth over the catchment, with a marked seasonal distribution. The actual magnitude of yield has little significance except in the context of storage reservoirs. What is important is the minimum flow, as discussed in Chapter 3. This determines whether storage or a supplementary source from groundwater is required or not. For livestock watering purposes small storage reservoirs may be needed, which will almost certainly fill up in the wet season provided there is not a complete failure of rain.
Since the use of small streams depends to a large extent on their dry-season flow, it is extremely valuable to gauge the streams during the dry season, and then to assess the reliability of that flow from an analysis of rainfall or streamflow records in the vicinity.
One characteristic of streams in semi-arid areas, which has a significant bearing on their use, is that they are commonly 'influent'. This means that they are discharging to groundwater through the stream bed, with the result that the quantity of flow diminishes noticeably downstream. This influent discharge is usually confined to the riparian zone which sustains a much denser vegetative cover. Thus, care must be taken in extrapolating streamflow measurements from one point to another, especially in the drier zones.
On the other hand, a stream may become 'effluent' as it crosses from a recharge to a discharge area, and springs in the bed or along the banks are usually reflected in more luxuriant vegetation, which can be easily picked out in aerial surveys or in colour imagery.
An occasional feature of pastoral lands is the existence of large perennial rivers which originate in higher, more humid regions. Seasonal variations in discharge are determined by the rainfall distribution in the humid parts of the catchment and by the geological controls on baseflow. Dry season flow, therefore, may be sufficient for the constant abstraction of water for livestock needs.
A characteristic feature of this type of river is the high sediment load it carries. This poses special problems at the site for abstraction, since any structure or diversion which causes a decrease in velocity in the river will cause deposition of sediment.
Such rivers are prone to change course during a flood, particularly if they are meandering over a wide, flat plain. It is not uncommon for pump houses, diversion weirs, bridges or other riverside structures to be either destroyed or left abandoned by a sudden change in course. Balancing reservoirs or barrages can assist with smoothing out the natural variations in flow, but their high cost is usually only justified for hydropower generation or multipurpose projects. However, livestock projects should take advantage of such schemes, since they afford a means of ensuring more reliable dry-season flows and more secure abstraction points.
Rivers originating wholly within range areas are rarely perennial and are always subject to the same problems of erratic seasonal flow. The wide range of discharge creates special problems in dealing with high floods, so that it is imperative that management schemes be introduced to lessen the magnitude of peak floods and to extend the duration of flow.
Natural lakes may be permanent or seasonal, fresh or saline depending on their water balance and many other limnological factors which are beyond the scope of this report. High evaporation rates and low inflow will clearly lead to the concentration of dissolved chemicals, and water quality will usually be a critical factor in deciding whether lakes can be used as a resource for livestock watering or not.
Permanence is another feature which determines whether or not lakes can be used. Many lakes persist during a run of wet years and then dry up completely when a prolonged dry spell occurs. There is often an abrupt change in water level, because a return to a wetter period is usually heralded by unusually high rainfall. A good example of this occurred during the season 1961/62 when the majority of the East African Rift Valley lakes rose to high levels, often 5 m above the previous year's maximum.
These fluctuations, both annual and seasonal, are a consequence of variability in rainfall. Unprecedented high levels of major lakes such as Lakes Victoria, Tanganyika and Malawi, during the 1960s and 1970s, led to speculation about the effect of man on the hydrological cycle - particularly with respect to changes in land use and the construction of artificial controls on outflow. Intensive studies (WMO, 1974; Kidd, 1983), however, have not been able to detect any evidence of man having other than a minor effect. There is overwhelming evidence, on the other hand, that a run of wetter-than-average years, such as occurs from time to time in a random sequence, can produce exceptionally high lake levels.
A corollary to the occurrence of wet periods in random sequences is the more disturbing possibility of a return to the drought periods which have been experienced at different times, and in different areas, since the beginning of this century when most records begin. The economic and social consequences of such droughts have already been felt in the Sahelian zone of West Africa. Where natural lakes are shallow and clearly dependent upon annual recharge, caution should be exercised in placing too great a reliance on their permanency.
4.2.1 The occurrence of groundwater
4.2.2 Prospecting for groundwater
4.2.3 Characteristics of the groundwater of semi-arid regions
The general principles behind the occurrence of groundwater have been discussed in Chapter 3. Aquifers can be of many different types, with varying characteristics. The geology of the pastoral areas of Africa is also varied and complex. The majority of existing boreholes in tropical Africa have been drilled without hydrogeological advice, by drillers armed only with an intimate knowledge of local conditions and many years of experience.
At the same time, many of the older existing boreholes have been located in regions of favourable hydrogeology, particularly recharge, because groundwater development first took place in regions of high agricultural productivity. It becomes increasingly difficult to locate sites as the recharge potential diminishes. Nowadays, there is a need for greater professional input, and there are legal requirements to be met in the construction and completion of boreholes. It is increasingly important to take into account potential yield, water chemistry and sound design and completion, in order to minimise both the capital costs and the maintenance costs of a borehole.
Most groundwater or hydrogeology textbooks deal in depth with the subjects of groundwater geology and exploration for groundwater (Todd, 1959; Davis and De Wiest, 1966; Freeze and Cherry, 1979). This section is intended only to give a brief background to the methods of exploiting groundwater resources. While the more expensive methods of exploitation (deep boreholes) should not be undertaken without the advice of a professional hydrogeologist, the less sophisticated methods (shallow wells and boreholes) can be applied often with a minimum knowledge of the hydrogeology of an area.
Freeze and Cherry (1979) summarise the main types of aquifers according to their geologic origin and their porosity. Dealing first with sedimentary rocks, the most abundant category is in fine-grained rocks such as shales, clays and silt-stones. These make up about 50% of all sedimentary rocks, and are characterised by having relatively high porosities but low permeabilities. They are commonly barriers to the movement of water, although they provide storage for large quantities of water which cannot easily be extracted.
Sandstones form the next largest group and, depending on the degree of cementation and packing of the individual grains, they can have porosities of 5 to 30%. Permeabilities are also very variable and seem to be a function mainly of grain size and degree of cementation.
Limestones (including dolomite) are another major group of consolidated sediments. Once again the fine-grained structure of limestones (excluding limestone breccias) tends to produce medium to high porosities but low permeabilities. The major characteristic of limestone, however, is the importance of fractures and solution openings, which are common features and can yield large quantities of water.
Undoubtedly the most important sedimentary formations, from the point of view of groundwater yield, are the unconsolidated or non-indurated sedimentary deposits such as alluvium, loess, dune sand, marine sands and clays, colluvial deposites and lacustrine clays and sands. Porosities range from 20% in coarse, poorly sorted alluvium to about 90% in soft muds and dry organic material (Todd, 1959). Porosities of between 25 and 65% are most common. Permeability and specific yield are dependent upon the shape, packing, size distribution and incipient cementation or clay coating of particles. Such sediments occur in alluvial basins and coastal margins. They are readily identifiable and indicate where high-yielding aquifers are most likely to occur. An understanding of the geomorphology or palaeogeomorphology will often help to identify buried valleys or deltaic fans, where the coarsest sediments are likely to be found. Inland drainage basins (which are common in arid climates) and lakes are areas where slack-water deposits such as silts and clays abound. Careful reconstruction of recent geological history may provide the necessary clues to the subsurface geological conditions, so that the highest yielding sediments can be located and traced.
Igneous and metamorphic rocks are usually regarded as giving the most unfavourable geological conditions for the occurrence of groundwater Precambrian basement complex rocks of this type cover large areas of tropical Africa. High-grade metamorphic rocks, often intensely folded and granitised, form the tectonic shield areas of the continent. They are impermeable in their fresh state with very low porosities (1 to 3%). Groundwater occurs in small fissures or fractures which are difficult to locate, and only where widespread faulting and fissuring has occurred will the yield be high enough for wells to be developed.
These rocks have been exposed to tropical weathering for millions of years, however, and the layer of unconsolidated material produced by in situ weathering forms a most important aquifer which has largely been ignored. Except where the weathered material has been stripped from the parent material by recent erosion or fluvial rejuvenation, or where rock resistant to erosion outcrops in erosion residuals, deep weathering produces a shallow aquifer, generally less than 30 m thick, which is of great importance as a source of rural water supply. This weathered zone will normally support a yield of 0.51/sec, which is adequate for a hand pump. Furthermore, provided that the areas where the bedrock is near the surface are avoided, such yields can be found over wide areas without the need for sophisticated siting techniques.
The exploitation of this aquifer has largely been neglected in the past, but recent work in Malawi (Chilton, 1983) has demonstrated its importance in rural areas. Boreholes drawing on this aquifer need only be 20 to 30 m deep, compared with 45 to 50 m in the past.
The final major group of rocks encountered in resource evaluation is the volcanic group. These have a very wide range of hydrogeological properties. Recent basalts may have very high transmissivities, whereas fine-grained tuffs or dikes usually have very low permeabilities. Vulcanicity is associated with the African Rift Valley and with tectonic plate movements on the eastern side of the continent. Here, the quaternary to recent volcanics form important aquifers, while the mesozoic volcanics have lower porosities and permeabilities. Boreholes in the mesozoic volcanics tend to be deep; the median value of boreholes in volcanic rocks in Kenya being 125 m (TAMS, 1979). Water chemistry of the deeper volcanic aquifers is frequently poor (for example in these Kenyan boreholes, high concentrations of fluoride are commonly encountered). The rapidly cooled surface basalt flows, on the other hand, yield high-quality water, often in large quantities.
Surface geological methods of prospecting for groundwater have been mentioned already. Analysis of available geological information (maps, cross-sections), air photographs and, more recently, satellite imagery, all contribute to assessing the location of aquifers. Hydrological techniques help in assessing the recharge potential of such aquifers. This is often a good guide to potential water chemistry, since the 'fossil' waters, which are not being flushed with fresh water, are the most likely areas of non-potable water in zones of otherwise favourable geology.
Hydrobotanical studies from aerial photographs can also be instructive. Phreatophytes take their water from shallow sources; halophytes, together with white surface encrustations, indicate shallow, brackish or saline groundwater; and xerophytes are plants resistant to drought, suggesting a considerable depth to the water table. Care must be taken when using the presence of phreatophytes as an indication of recoverable groundwater. Plants are able to abstract water from impermeable, fine-grained materials which would not normally contribute water to a well or borehole.
Geophysical methods cover a range of techniques for interpreting the subsurface geology. They are usually much more expensive than geological or hydrological reconnaissance surveys, and are only used where the yield is important, or the economics of utilising groundwater justify the cost of locating the source. They range from simple electrical resistance traverses with different electrode configurations to seismic refraction techniques and magnetic and gravity surveys.
All geophysical methods have limitations where the subsurface geology is complex, and they are usually supplemented by the analysis of existing borehole records or the results of test drilling. Ideally, both drilling and geophysics should be used, but economic constraints dictate the extent to which subsurface methods can be used at all. Clearly, if any drilling takes place, careful logging with depth will yield valuable direct information. Once again, logging can range from simple lithological logging to electrical logging (spontaneous potential and resistivity) and radiation logging (gamma and neutron). Bailing tests will also give a coarse indication of yield with depth.
In the context of pastoral areas geophysical techniques will seldom be economic. However, comprehensive reconnaissance surveys, backed by local information on the location of springs and the depth and reliability of shallow wells, can often produce sufficient information for a preliminary drilling programme. As information accumulates from the logging of these first boreholes, so the picture of subsurface conditions will become clearer and the feasibility of a second phase can be assessed.
As the aridity of the climate increases, certain characteristic features appear which can have a dominating influence on the development of the groundwater resource. A general decrease in the volume of streamflow can often lead to the inability of a river to maintain its course over volcanic or tectonic obstructions (Todd, 1959). Enclosed drainage basins are common, therefore, filled with fine lacustrine deposits mixed with saline residues. Fresh water is usually found only on the margins of such basins, or in deeper lying aquifers which are not affected by the upward movement or surface concentration of salts in response to high evaporation rates. Recharge is commonly through channel bottoms or through the vegetated parts of isolated hills. Small recent volcanic features act as collecting catchments for rainfall, particularly where craters still exist.
In the African context, a knowledge of the geomorphology of arid regions will often prevent abortive efforts to locate groundwater in impermeable bedrock. The formation of pediments at the base of erosion residuals, and rock pavements where wind erosion is significant, should be borne in mind during the initial reconnaissance. Pediments are often covered with a thin mantle of colluvium which makes them resemble alluvial fans. Apparently dense, quasi-parallel drainage networks on extensive peneplains indicate impermeable bedrock near the surface, where both streamflows and slope are insufficient to cause concentration of flow into fewer, well defined channels. Rejuvenation of drainage systems by recent tectonic activity can also cause the stripping of the weathered mantle from basement rocks. All the above indications can be used to avoid areas of shallow bedrock where, even if fissures and fractures abound, storage and recharge are likely to be insufficient for even small human or animal communities.
Another feature of the semi-arid regions is the existence of large 'sand rivers'. Either high rates of geologic erosion or accelerated erosion produce large quantities of coarse material which have a potential for development.
Under natural conditions available water is limited, and this acts as a control on the degradation of forage within easy reach of the water course. This situation can be easily upset if artificial methods are introduced to prolong the availability of water. The balance is a delicate one, and development of permanent watering points can have a negative effect if forage availability and water supply are not considered together. With that factor in mind, the exploitation of the available resource can now be discussed, leaving the questions of forage and water needs, and the organisation and management of water supplies, to the companion reports in this series.
